Matched adjacent normal and tumor tissues from 45 CRC patients and the HCC samples were collected at the Johns Hopkins University (JHU) as per IRB approved protocol. Formalin-fixed paraffin-embedded (FFPE) tumor tissue sections (5 μm) were procured from all 45 CRC patients for immunohistochemical (IHC) analysis. All the subsequent analysis of tissue samples, IHC staining was performed at JHU. Detailed patient information is shown in Table I.

We obtained 2 human non-tumorigenic colonic cell lines; CCD841 and FHC and two hepatocellular carcinoma (HCC) lines H2pG2 and Hep3B from ATCC. All cell lines were cultured in ATCC recommended medium. The cell lines were periodically chacked for mycoplasma contamination using Mycoplasma Detection kit from Sigma (MP-0025). The tissue culture media and reagents were purchased either from ATCC or Invitrogen.

Genomic DNA was extracted from the microdissected tumor and normal tissue samples and cell lines as described (12). The list of primers is given in Table II. The PCR was carried out in a 50 μl reaction volume with the conditions as follows: 95°C for 5 min, 1 cycle; 94°C for 30 sec, 55°C for 30 sec, 72°C for 7 min, 40 cycles; and final extension at 72°C for 10 min. All amplified products were analyzed on 1% agarose gels, cut, purified and sequenced using the ABI Big Dye Cycle Sequencing kit (Applied Biosystems). All experiments were carried out at JHU.

TaqMan™ (Fluorogenic 5′ nuclease) assay was used for typing our candidate SNP (rs1048575). The forward and reverse primers and respective allele-specific probes were obtained from Applied Biosystems (Applied Biosystems). PCR was conducted in ABI 9700 thermocycler, and the end-point results were auto-scored using the ABI 7900HT Sequence Detection System (SDS; version 2.3). To ensure reliability of genotyping data, each of the samples were amplified and genotyped in triplicate along with positive and negative control templates. Additionally, all the samples were sequenced unidirectionally, and there was 100% concordance between sequencing and genotyping data obtained through the TaqMan genotyping assay. All experiments were carried out at JHU.

To study the expression of PIGK protein, we used FFPE tissue samples collected from 45 samples described in Table I. Several serial sections (5 μm) from different regions of the tumors were stained using rabbit anti-human PIGK antibody (Abcam #38445) for determining protein expression. Staining and evaluation was performed as described earlier (13,14). For comparison, all sections were processed in parallel. All experiments were carried out at JHU.

The wild-type (wt, C/C) human PIGK gene was subcloned into SalI and NotI sites of the phosphorylated cytomegalovirus (pCMV)/myc/ER plasmid (Invitrogen) to express the exogenous protein in the endoplasmic reticulum (ER) as PIGK is localized and functions in the ER (10). The resultant plasmids were resequenced using the ABI BigDye cycle sequencing kit (Applied Biosystems) for verification of the insert sequences. An empty pCMV/myc/ER and a Green Fluorescence Protein (GFP) expressing pCMV/myc/ER-GFP plasmids were also used as controls. In transfections, HepG2 and Hep3B cells were transiently transfected with wt-PIGK plasmid in the presence of the FuGene 6 transfection reagent (Roche Applied Science). An empty pCMV/myc/ER vector was used for mock transfection. We also used GFP expressing pCMV/myc/ER plasmid vector as a control. Cells were analyzed 24 h post transfection. Positive cell clones were confirmed to express wt-PIGK by immunofluorescence analysis using the fusion protein tag anti-myc-FITC antibody (Invitrogen).

Total RNA was extracted from wt-PIGK transfected and non-transfected Hep3B cells after 24 h using Qiazol reagent following the manufacturer’s instruction (Qiagen). Conventional RT-PCR was carried out in 50 μl reaction volume as described earlier (14). The primers used for PIGK were forward 5′-CCACTTTGCCTTTCTCTCCA-3′; reverse 5′-GCCCCATTCAGATCCTCTTC-3′.

χ² or Fisher’s exact tests were used as appropriate. Student’s t-test or non-parametric Wilcoxon rank sum test was also used as appropriate. All p-values are two-sided and all confidence intervals are at the 95% level. Computations for all the analysis were performed using the Statistical Analysis System (SAS).

In an earlier study, we observed a low expression of PIGK protein in tumors compared to paired normal tissues of some CRC patients (11). We hypothesized that such low expression could be due to mutation in the coding regions of the PIGK gene. Therefore in this pilot study, we examined all 10 exons of the PIGK gene by direct sequencing in 45 primary CRC tumors. No mutations were detected in the coding regions, however, we discovered a SNP (rs1048575; C/C→C/G or G/G) in the 3′UTR of the PIGK gene in 67% (30/45) of the CRC patients (Table I and Fig. 1). Schematic representation of the location of the ancestral and altered allele in PIGK is shown in Fig. 2. This observation was further confirmed by genotyping the SNP-1048575 in all the above samples (Table I, Fig. 1B). Of the 45 CRC patients, 26 were of Jewish and 17 were of Arabian origin. The ethnicity of the remainder of the 3 patients was unknown. Eighty-five percent (22/26) of the Jewish patients showed the altered alleles (C/G or G/G) whereas 47% (8/17) of the Arabian patient populations was detected with the altered alleles (Table I, Fig. 1A and B). The number of Jewish patients with the altered alleles was significantly higher (Student’s t-test; p=0.03) compared to the Arabian patient populations.

To examine whether there was any correlation between the altered alleles (C/G, G/G) and PIGK protein expression in the CRC patients, we performed IHC on FFPE tissue sections from all 45 CRC patients. The expression of PIGK protein was significantly lower (p<0.002) in all of the patients with altered genotype (C/G or G/G; 30/45, 67%) compared to the ancestral genotype (C/C; 15/45, 33%) (Table I and Fig. 1C).

We also determined the SNP-1048575 allele distribution and corresponding PIGK expression in a couple of transformed colon cell lines. One colon line (CCD841) exhibited the ancestral alleles (C/C), whereas the other line (FHC) exhibited the altered alleles (C/G) (Fig. 3A). Immunohistochemical analysis also revealed a significantly lower (Student’s t-test, p<0.002) expression of the PIGK protein in the FHC-colon line carrying the altered alleles (C/G) compared to the CCD-841 cell line with the ancestral alleles (C/C) (Fig. 3B).

Similar to the CRC patients, our earlier study also demonstrated a low expression of PIGK in some primary HCC tumors (11). Therefore, we also examined 5 HCC patients and two HCC cell lines (Hep3B and HepG2) for PIGK genotype (SNP-1048575) and corresponding protein expression. We observed altered alleles (C/G or G/G) (Fig. 4A) and corresponding low PIGK protein expression (Fig. 4B) in 4 out of 5 (80%) primary HCC tumors. Among the HCC cell lines, HepG2 line exhibited ancestral C/C alleles, whereas Hep3B showed altered C/G alleles (Fig. 4C, upper panel). Similar to the HCC patients, Hep3B line with the altered alleles (C/G) exhibited significantly low (Student’s t-test, P<0.002) PIGK protein expression compared to the Hep3B line carrying the ancestral (C/C) alleles (Fig. 4C, lower panel).

To examine the exogenous PIGK protein expression status, we transiently transfected both HepG2 (C/C alleles) and Hep3B (C/G alleles) cell lines with wt-PIGK constructs. As shown in Fig. 4D (upper panel), we could detect exogenously expressed PIGK protein in HepG2 (C/C) cells, but no PIGK expression was detectable in Hep3B (C/G) cells. The control GFP protein was detectable in both cell lines using the same expression cassette. We also examined the mRNA expression level of the exogenously expressed wt-PIGK in Hep3B (C/G) cells. As shown in Fig. 4D (lower panel), no mRNA expression was detected in Hep3B cells.